Virulence Determinants and Methicillin Resistance in Biofilm-Forming Staphylococcus aureus from Various Food Sources in Bangladesh

The eradication of staphylococcal infections has become more difficult due to the development of antibiotic resistance and virulence in biofilm-forming Staphylococcus aureus. The presence of the life-threatening zoonotic pathogen, methicillin-resistant S. aureus (MRSA), in foods indicates a public health issue. This study, therefore, aimed to determine virulence factors and methicillin resistance in biofilm-forming S. aureus isolates from different foods and food handlers. A total of 100 PCR-positive S. aureus isolates (97 biofilm formers and three non-biofilm formers) were screened using the disk diffusion method and PCR assay. By PCR, genes encoding virulence factors, e.g., enterotoxin (sea, 30%, 95% CI: 21.90–39.59%), toxic shock syndrome toxin (tst, 20%, 95% CI: 13.34–28.88%), and Panton–Valentine leukocidin toxin (PVL, 15%, 95% CI: 9.31–23.28%), were detected in the S. aureus isolates. By the disk diffusion method, 100% (95% CI: 96.30–100.00%) of S. aureus isolates were phenotypically MRSA in nature, showing 100% resistance to oxacillin and cefoxitin. Moreover, the methicillin-resistant gene mecA was found in 61 (61%, 95% CI: 51.20–69.98%) MRSA isolates. Furthermore, all the S. aureus isolates were phenotypically resistant to ampicillin and penicillin, 30% to erythromycin, and 11% to gentamycin. Among them, 51% (95% CI: 41.35–60.58%) of S. aureus isolates were phenotypically multidrug-resistant in nature, and the multiple antibiotic resistance index varied from 0.33 to 0.55. Genes encoding resistance to beta-lactams (blaZ, 100%, 95% CI: 96.30–100.00%) and tetracyclines (tetA and tetC, 3%, 95% CI: 0.82–8.45%) were found positive in the S. aureus isolates. Genes encoding virulence determinants and MRSA were significantly (p < 0.05) higher in strong biofilm-forming S. aureus than in moderate and non-biofilm-forming isolates. To our knowledge, this is the first study in Bangladesh to incorporate preliminary data on the occurrence of virulence determinants and methicillin resistance, including resistance to clinically important antibiotics, in biofilm-forming S. aureus isolates from different foods and food handlers in Bangladesh, emphasizing a potential threat to human health.

Sample-wise, S. aureus isolates detected from ready-to-eat foods showed the highest occurrence of different virulence genes compared with other samples (Supplementary Table S1).
Bivariate analysis revealed strong and positive significant correlations between virulence genes sea and tst (Pearson correlation coefficient, ρ = 0.600), sea and PVL (ρ = 0.458), and tst and PVL (ρ = 0.420). Table 2 shows the overall correlation outcomes between the virulence genes of S. aureus isolates.

Antibiogram Profiles of Biofilm-Forming S. aureus
By disk diffusion assay, all the isolated S. aureus exhibited resistance to oxacillin and cefoxitin (100/100, 95% CI: 96.30-100%), which indicates that all the 100 isolates were phenotypically MRSA in nature. In addition, resistance to ampicillin and penicillin was found in every single isolate. Resistance to erythromycin and gentamicin was estimated at 30% and 11%, respectively. Sensitivity to chloramphenicol, ciprofloxacin, co-trimoxazole, tetracycline, and azithromycin was estimated at 98%, 98%, 87%, 86%, and 85%, respectively. Figure 1 shows the overall antibiotic susceptibility profiles of 100 S. aureus isolates. Samplewise phenotypic antibiotic resistance patterns are documented in Supplementary Table S1 and Figure 1. Moreover, there was no statistically significant correlation between any of the two antibiotics' resistance in S. aureus isolates (Supplementary Table S2).

Association of Antibiotic Resistance Patterns with Biofilm-Forming S. aureus
The resistance patterns of all the antibiotics except oxacillin, ampicillin, penicillin, and cefoxitin (which showed 100% resistance in all levels of biofilm-forming isolates) were higher in biofilm-forming S. aureus isolates than in non-biofilm-forming isolates. Moreover, the occurrence of erythromycin resistance patterns was significantly (p < 0.05) higher in strong biofilm-formers (55%), compared with intermediate (24.68%) and nonbiofilm-forming (0%) isolates. Strong biofilm-forming S. aureus isolates showed higher resistance to other antibiotics (except chloramphenicol and ciprofloxacin) compared with moderate and non-biofilm-producing S. aureus isolates, but there were no significant variations (p > 0.05) ( Table 3).

Association of Antibiotic Resistance Patterns with Biofilm-Forming S. aureus
The resistance patterns of all the antibiotics except oxacillin, ampicillin, penicillin, and cefoxitin (which showed 100% resistance in all levels of biofilm-forming isolates) were higher in biofilm-forming S. aureus isolates than in non-biofilm-forming isolates. Moreover, the occurrence of erythromycin resistance patterns was significantly (p < 0.05) higher in strong biofilmformers (55%), compared with intermediate (24.68%) and non-biofilm-forming (0%) isolates. Strong biofilm-forming S. aureus isolates showed higher resistance to other antibiotics (except chloramphenicol and ciprofloxacin) compared with moderate and non-biofilm-producing S. aureus isolates, but there were no significant variations (p > 0.05) ( Table 3). Values with different superscripts differ significantly (p < 0.05) within the variable under assessment, CI = Confidence interval, NA= Not applied, n = number of isolates, C = Chloramphenicol, E = Erythromycin, COT = Co-trimoxazole, CIP = Ciprofloxacin, AZM = Azithromycin, GEN = Gentamicin, OX = Oxacillin, AMP = Ampicillin, TE = Tetracycline, P = Penicillin, and CX = Cefoxitin.
Values with different superscripts differ significantly (p < 0.05) within the variable under assessment, CI = Confidence interval, NA = Not applied, n = number of isolates.
Moreover, all the biofilm-forming S. aureus isolates were found positive for at least one antibiotic resistance gene (Supplementary Table S1). The tetracycline resistance genes tetA and tetC were detected in 3% (95% CI: 0.82-8.45%) of S. aureus isolates, and the beta-lactam gene blaZ was in 100% of S. aureus isolates. However, there was not a significant difference (p > 0.05) between the level of biofilm formation and the occurrence of these resistance genes. No isolates were found to be positive for tetC (Table 5). Sample-wise data on different antibiotic resistance genes are given in Supplementary Table S1.

Discussion
Foods, especially ready-to-eat foods, are becoming incredibly popular, with an increased number of restaurants and street vendors around the globe, most notably in Bangladesh. Other food sources such as milk, meat, fish, and eggs always have a high demand among all classes of people. However, foods contaminated with S. aureus have the potential to cause food poisoning, generating serious public health risks. In addition, biofilm formation in S. aureus is responsible for different persistent or chronic staphylococcal infections. In this study, we present the first-time detection of virulence determinants and methicillin resistance in biofilm-forming S. aureus isolated from different foods and human hand swab samples in Bangladesh. Islam et al. [22] conducted almost-similar types of research in Bangladesh. They reported virulence factors and antibiotic resistance only in S. aureus isolates from food sources, but they did not focus on biofilm. In addition, we showed variations in the occurrence of virulence and antibiotic resistance with different degrees of biofilm formation in S. aureus isolates, but they did not.
Staphylococcal enterotoxins (sea, seb, and others) produced by S. aureus are directly associated with staphylococcal food poisoning [30]. The higher prevalence of the sea gene than the seb gene in S. aureus isolates detected in the present study is not unusual because the isolates with SEA-type toxins cause the most staphylococcal infections and outbreaks, followed by isolates with other staphylococcal enterotoxin-related infections [24]. Detecting genes encoding staphylococcal enterotoxins in the isolated staphylococcal species indicates a serious public health concern since these toxins have resistance activity against high temperatures and can even retain their biological properties in milk during pasteurization [31]. The tst gene is related mainly to human S. aureus isolates [24]. The presence of the tst gene in food and human hand swab samples suggests that this type of gene could be transferred from humans to animals and vice versa via the food chain. Another virulence gene, PVL, is a pertinent S. aureus virulence gene that is attributed mostly to community-acquired infections [32]. The detection of the PVL gene in our study suggests that food can be contaminated by this virulence gene, and these spoiled foods could perhaps constitute a source of community-acquired infections. In addition, virulence genes were found to be significantly higher in strong biofilm-forming S. aureus isolates. This indicates that as the degree of biofilm formation in S. aureus isolates increases, so does their ability to develop infections.
Treatment of staphylococcal infections relies mostly on antibiotic therapy; however, it frequently fails due to their resistance to antibiotics. In addition, the presence of MRSA in foods and on food handlers' hands raises serious public health implications. In this study, all of the isolates were phenotypically MRSA in nature, indicating a critical threat to consumers by limiting the treatment options. Animal-originated food and food products can be contaminated with MRSA from infected animals, and processed food can be contaminated by infected vendors during food processing. The detection of MRSA in food and human hand swab samples indicates a serious issue for human health because it could be transmitted to humans via the food supply chain, causing staphylococcal infection. In addition, MRSA strains are responsible for severe morbidity and mortality in hospitals and even in healthy individuals [33]. Previously in Bangladesh, Islam et al. [24] reported that 25.7% (9/35) of S. aureus isolates originating from food samples were MRSA in nature, which is relatively lower than our findings. However, Saber et al. [17] recorded a similar prevalence rate, detecting MRSA in 100% of food samples in Egypt. In addition, previous studies found MRSA in different foods and human hand samples with lower and higher occurrence rates, i.e., 57.1% in food handlers in Brazil [34]; 10.1% in retail ready-to-eat foods [35] and 8.4% in retail aquatic products [27] in China; 12.94% in retail chicken meat and eggs in Nepal [36]; 42.3% in street-vended foods in India [37]; and 22.6% in ready-to-eat meat sandwiches in Egypt [38]. It is possible that different strategies for the distribution and usage of antibiotics in humans and animals account for the varying rates of MRSA detection in food samples. Moreover, S. aureus isolates detected in this study were highly resistant to ampicillin (100%), penicillin (100), and erythromycin (30%), but highly sensitive to chloramphenicol.
In total, 51% of the isolates detected in this study were MDR in nature, and the MAR index of the isolates varied from 0.33 to 0.55, demonstrating that antibiotics were haphaz-ardly used at the source of the contamination. According to Krumperman [39], isolates with a MAR index of more than 0.2 were thought to have come from a high-risk source of contamination where antibiotics are commonly used. In Bangladesh, antibiotics are readily available in the markets, and there are no proper established policies for their distribution and usage. Therefore, people use antibiotics without any prescriptions and/or consultation with human physicians or veterinarians. The overuse and misuse of antimicrobial agents in both humans and animals might be responsible for the development of MDR S. aureus isolates in different food sources.
The association analysis between the degree of biofilm formation and phenotypic antibiotic resistance revealed that most antibiotics (except ampicillin, penicillin, oxacillin, and cefoxitin) showed resistance only to biofilm-forming S. aureus isolates. Antibiotic resistance increased with the level of biofilm formation, including the determination of significantly higher resistance to erythromycin in strong biofilm-forming S. aureus isolates. Antimicrobial resistance in S. aureus has been reported to increase to 1000 times that of planktonic cells in the presence of biofilm [40]. Various factors explaining the unprecedented resistance of biofilm-forming S. aureus and other bacteria to antibiotics include: (1) lower or decreased metabolic and growth rates of biofilm-formers, which may render them intrinsically less sensitive to antibiotics; (2) the structure of the biofilm-EPS (extracellular polymeric substances) matrix that assists biofilm cells to reduce the access of antibiotics to regions of the biofilm; and (3) the distinct physiological characteristics of biofilm cells that help to express MDR efflux pumps and stress-response regulons for developing antibiotic resistance [41].
In our study, 61% (61/100) of biofilm-forming S. aureus isolates harbored the methicillin resistance gene, mecA. The standard gold method for detecting MRSA isolates is mecA detection via PCR [42]; however, in our study, the mecA gene was absent in 39% of S. aureus isolates that were phenotypically MRSA. The inconsistency of the correlation between phenotypic and genotypic resistance of MRSA might be due to the mutation of genes that result in non-functional proteins and the dearth of gene expression [43]. In addition, the absence of the mecA gene in the MRSA isolates might be due to the detection methods we used, or those MRSA isolates could harbor other methicillin resistance genes, such as mecB, mecC, or mecD [42]. Our findings suggest the possible presence of other intrinsic and extrinsic factors having the ability to compete with the mecA gene for developing MRSA. Furthermore, the resistance gene mecA was significantly higher in strong biofilm-forming S. aureus isolates compared with moderate and non-biofilm producers. The genotypic detection of MRSA in biofilm-forming S. aureus isolates from foods suggests a serious threat to human health because these resistance gene-containing isolates could easily be transferred to humans via the food supply chain. In addition, it would be challenging to manage these organisms clinically because of their biofilm-forming activities.
Penicillin-resistant S. aureus isolates could be interpreted as beta-lactamase resistance. Detecting the blaZ gene using PCR is also necessary to determine the occurrence of betalactamase-producing isolates [37]. In this study, blaZ was found positive in all the S. aureus isolates, which is relatively higher than the findings (69.23%) from a previous study [37]. Tetracycline resistance genes tetA and tetC were also detected in biofilm-forming S. aureus isolates. The detection of antibiotic resistance genes in our S. aureus isolates from foods and human hand swab samples suggests that these resistance genes might be transferred to other bacteria via horizontal transmission. We found these resistance genes in both foods and handlers' hand swab samples, indicating a high-risk S. aureus contamination developed by the poor hygienic condition of the sampling sites.

Selection of S. aureus Isolates
Staphylococcus aureus (n = 100, biofilm producers = 97, and non-producers = 3) strains and their data were obtained from our previous study on detecting biofilm-producing S. aureus from different foods (raw milk, egg surface, chicken muscle, fish, and ready-to-eat foods) and humans' hand swab samples [44]. Originally, S. aureus isolates were identified by culturing on mannitol salt (MS) agar plates, applying different bacteriological analytical methods (Gram's staining, glucose, and mannitol utilization tests, coagulase test, catalase test, and Voges-Proskauer tests), and finally, employing the polymerase chain reaction (PCR) test targeting the nuc gene [44]. The biofilm-producing ability of S. aureus isolates was evaluated by qualitative (Congo red agar plate test), quantitative (crystal violet microtiter plate test), and genotypic (PCR) assays [44]. All the data related to biofilm-forming S. aureus are documented in Supplementary Table S1.

Molecular Detection of Virulence Factors
PCR-confirmed S. aureus isolates were subjected to a simplex PCR for the detection of virulence factors, namely staphylococcal enterotoxin A (sea), staphylococcal enterotoxin B (seb), tst, and PVL (Table 6). The DNA for PCR was extracted from pure cultures of S. aureus using the boiling technique [51,52]. The genomic DNA was amplified using a PCR thermal cycler (ASTEC, Fukuoka, Japan). The PCR mixture was prepared following the previous study [44], and the PCR conditions were set following the previous studies mentioned in Table 6. The PCR products that had been amplified were then run through a gel electrophoresis machine (Nippon Genetics, Tokyo, Japan) using 1.5% agarose (Invitrogen, Waltham, MA, USA). After completing the gel run, the products were stained using ethidium bromide (HiMedia, Maharashtra, India) and checked for their expected amplicon sizes using an ultraviolet trans-illuminator (Biometra, Göttingen, Germany). A 100-bp DNA ladder (Promega, Madison, WI, USA) was used to compare the sizes of the bands.
Isolates showing resistance to methicillin and cefoxitin were considered phenotypically MRSA in nature. All the antibiotic disks were collected from two different manufacturers (HiMedia, Maharashtra, India, and Oxoid, Hampshire, UK), and all the susceptibility tests were done three times to confirm the exact results. MDR isolates were those that showed resistance to more than two antimicrobial categories [55]. Moreover, the index of multiple antibiotic resistance (MAR) was enumerated by the previously described formula [39]: The number of antimicrobials to which a given S. aureus strain is resistant The sum of antibiotics to which an isolate was subjected

Molecular Detection of MRSA with other Antibiotic Resistance Genes
All the S. aureus isolates were subjected to PCR to detect their genotypic antibiotic resistance profiles. The molecular detection of MRSA was performed by detecting the methicillin resistance gene (mecA) using a PCR assay. Moreover, the genes associated with resistance to beta-lactams (blaZ) and tetracyclines (tetA, tetB, and tetC) were also tested using PCR ( Table 6). The same amplification method was used to detect antibiotic resistance genes in S. aureus isolates as was used to detect virulence genes.

Statistical Analysis
Excel 365 (Microsoft/Office 365, Redmond, WA, USA) was used to incorporate data; GraphPad Prism (Prism 8.4.2, San Diego, CA, USA) and the Statistical Package for Social Science (IBM SPSS 25.0, Chicago, IL, USA) were used to analyze data.

Descriptive Analysis
Descriptive analysis was employed to calculate the prevalence of different variables. To estimate the prevalence, a binomial 95% confidence interval (CI) was enumerated by GraphPad Prism using a previous method [56]. Moreover, using SPSS, the variations in the occurrence of virulence and antibiotic resistance with the occurrence of different degrees of biofilm formation in S. aureus isolates were determined by the chi-square test for relatedness (with the Z-test for proportion) with a ≤ 0.05 significance p-value.

Bivariate Analysis
Using SPSS, the Pearson correlation coefficients were calculated to check whether any of the two antibiotics resistant (phenotypically) to S. aureus were correlated (statistically significant). Only those antibiotics that did not show constant resistance against S. aureus isolates were analyzed for the bivariate analysis. Moreover, a correlation between any two virulence genes was also determined. The significance level for the correlation was selected at p ≤ 0.05.

Conclusions
As stated, our study detected virulence determinants and methicillin resistance in biofilmforming S. aureus isolates sourced from different foods and human hand swab samples for the first time in Bangladesh. This study shows a high prevalence rate of genes encoding virulence factors and methicillin resistance in S. aureus isolates. The S. aureus isolates also showed resistance to several clinically important antibiotics, which demonstrates a potential public health concern by transferring to humans from foods via the food supply chain. In addition, as revealed in the present study, their presence in foods and food handlers indicates that foods such as raw milk, chicken meat, ready-to-eat foods, fish, and eggs could be a possible source of virulent MRSA with significant clinical relevance. Standardized surveillance and monitoring programs, combined with an organized educational awareness campaign on AMR and good hygiene practices, should be implemented throughout the food production and supply chain to reduce the colonization and dissemination of virulent MRSA and biofilm strains and to guarantee the safety of foods.